Create PercentFormatter

[sffc]

We do not currently have percentage formatting. We will need it for ECMA-402 compatibility.

We can do this in the scope of measurement unit formatting, but I think it would be useful to have a standalone percentage formatter since this is how it worked in previous versions of ECMA-402.

As a first step, we can make a data struct, probably in the icu_dimension crate.

[sffc]

A few tips to get started.

First, review the tutorial here: https://github.com/unicode-org/icu4x/blob/main/docs/process/writing_a_new_data_struct.md

Specifically for the percent formatter: you can create a data struct in experimental/dimension/src/provider.rs, something like this:

#[icu_provider::data_struct(PercentEssentialsV1Marker = "percent/essentials@1")]
#[derive(Default, Clone, PartialEq, Debug)]
#[cfg_attr(
    feature = "datagen",
    derive(serde::Serialize, databake::Bake),
    databake(path = icu_dimension::provider),
)]
#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
pub struct PercentEssentialsV1<'data> {
    /// Percent format with plus sign
    pub plus_pattern: Cow<'data, str>,
    /// Percent format with no sign
    pub zero_pattern: Cow<'data, str>,
    /// Percent format with minus sign
    pub minus_pattern: Cow<'data, str>,
    /// Index of the number placeholder within the plus pattern
    pub plus_index: u8,
    /// Index of the number placeholder within the no-sign pattern
    pub zero_index: u8,
    /// Index of the number placeholder within the minus pattern
    pub minus_index: u8,
}

We can make that a little bit more efficiently packed but I like to start with the clear and simple model.

Next, create provider/datagen/src/transform/cldr/decimal/percent.rs: this is where you will put your transform code. You need to have at least these impls:

impl DataProvider<PercentEssentialsV1Marker> for crate::DatagenProvider
impl IterableDataProviderInternal<PercentEssentialsV1Marker> for crate::DatagenProvider

You can see the other impls for inspiration.

Then, to plug it together, add your new data key in at least two places:

1. provider/datagen/src/registry.rs
2. pub const KEYS in experimental/dimension/src/provider.rs

With this, you should be able to run cargo make testdata at the project root and see the JSON data being generated. Once you get to this point, it is ready for a code review.